Table 3 Parameters of classifier and regression for machine learning methods.
From: Predicting loss aversion behavior with machine-learning methods
Methods | Criterion | n estimators | Splitter | Max depth | Min samples split | Min samples leaf | Min weight fraction leaf | Max features | Random state | Max Leaf Nodes | Min İmpurity Decrease | Min İmputrity Split | Bootsrap | Class weight | ccp_alpha |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Decision Tree Classifier | gini | – | Best | None | 2 | 1 | 0 | None | None | None | 0 | 0 | – | None | 0 |
Decision Tree Regressor | mse | – | Best | None | 2 | 1 | 0 | None | None | None | 0 | 0 | – | None | 0 |
Random Forest Classifier | gini | 100 | – | None | 2 | 1 | 0 | Auto | None | None | 0 | 0 | True | None | 0 |
Random Forest Regressor | mse | 100 | – | None | 2 | 1 | 0 | Auto | None | None | 0 | 0 | True | – | 0 |
Methods* | C | Epsilon | Kernel | Degree | Gamma | Coef0 | Shrinking | Probability | Tol | Cache size | Class weight | Verbose | Max İter | Decision function shape | Break ties | Random state |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Kernel SVC | 1 | – | Linear | 3 | Scale | 0 | True | False | 1e−3 | 200 | None | False | −1 | ovr | False | None |
Kernel SVR | 1e3 | 0.4 | rbf | 3 | 1e-1 | 0 | True | – | 0.001 | 200 | – | False | −1 | – | – | – |
Methodsa | n neighbors | Weights | Algorithm | Leaf size | p | Metric | Metric params | n jobs |
|---|---|---|---|---|---|---|---|---|
k-NN Classifier | 5 | Uniform | Auto | 30 | 2 | Minkowski | None | None |
k-NN Regressor | 5 | Distance | Auto | 30 | Minkowski | None | None |
